Prober controlling device and prober controlling method

Abstract

To provide a prober control device and a prober control method that can more accurately predict the tip position of a probe needle.SOLUTION: A prober controlling device that brings a probe needles (35) into contact with an inspection target for electrical characteristics includes: an input data acquisition unit (42) that acquires input data including temperature data of at least one of a probe card (26) and a card holder (25); a prediction unit (46) that predicts a tip position of the probe needle on the basis of the input data acquired by the input data acquisition unit, using a prediction model (47) that receives the input data as input and outputs the tip position of the probe needle; and a determination unit (50) that determines whether or not to execute prediction by the prediction unit on the basis of the input data used as training data for machine learning of the prediction model and the input data acquired by the input data acquisition unit, before the prediction by the prediction unit.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present invention relates to a prober control device for a prober used to inspect the electrical characteristics of semiconductor chips formed on a wafer, a prober control method, and the prober itself. [Background technology]

[0002] A plurality of semiconductor chips, each having the same electrical element circuit, are formed on the surface of a wafer. Before being individually cut into semiconductor chips by a dicer, the electrical characteristics of each semiconductor chip are inspected by a wafer test system. This wafer test system includes a prober and a tester (see Patent Documents 1 to 4).

[0003] The prober holds the wafer on the wafer chuck and moves a probe card with probe needles relative to the wafer chuck to bring the probe needles into electrical contact with the electrode pads of the semiconductor chip. The tester supplies various test signals to the semiconductor chip via terminals connected to the probe needles, and receives and analyzes signals output from the semiconductor chip to test whether the semiconductor chip operates normally.

[0004] Semiconductor chips are used in a wide range of applications and are operated over a wide temperature range. Therefore, semiconductor chip inspections must be performed at, for example, room temperature (normal temperature), high temperatures, and low temperatures. For this reason, the wafer chuck of a prober is provided with a temperature adjustment unit, such as a heater mechanism, a chiller mechanism, or a heat pump mechanism, which heats or cools the wafer held on the wafer chuck.

[0005] At this time, the temperature of each part of the prober other than the wafer chuck also gradually changes to approach the temperature of the wafer chuck. As a result, each part deforms due to thermal expansion caused by heating or contraction caused by cooling, and this deformation also changes the relative position of the probe needle and the semiconductor chip. As a result, when the probe needle and the wafer are moved relative to each other to test the semiconductor chip, a probing error may occur in which the probe needle does not properly contact the semiconductor chip.

[0006] Therefore, Patent Document 1 discloses a prober that attaches a temperature sensor to a probe card having probe needles and corrects the height position of a wafer chuck when the probe needles are brought into contact with a semiconductor chip based on the measurement results of the temperature sensor. In the prober described in Patent Document 1, the relationship between the temperature of the probe card and the amount of displacement of the probe needles in the height direction is determined in advance, so that the amount of correction for the height position of the wafer chuck can be determined from the measurement results of the temperature sensor.

[0007] Patent Document 2 discloses a prober in which temperature sensors are provided on a probe card and an X-direction moving stage, and the probe needles are brought into contact with the semiconductor chips while the temperature of a predetermined portion of the prober is stabilized based on the measurement results of the temperature sensors. The prober described in Patent Document 2 can shorten the preheating time for the wafer, probe card, etc.

[0008] Patent Document 3 discloses a prober that has temperature sensors attached to a wafer chuck, a card holder that holds a probe card, and a head stage that holds the card holder, and that corrects the contact position between a probe needle and a semiconductor chip based on the measurement results of each temperature sensor. The prober in Patent Document 3 generates a prediction model that shows the change in probe needle position due to changes in each temperature by determining in advance the relationship between the temperatures of the wafer chuck and the card holder and the position of the probe needle. This allows the prober in Patent Document 3 to correct the contact position between the probe needle and the semiconductor chip by referring to the prediction model based on the temperature measurement results of each temperature sensor.

[0009] Patent Document 4 discloses a prober that measures the temperatures of both a probe card and a card holder, and predicts the tip position of a probe needle based on the temperature measurement results by referring to a prediction model that shows the relationship between the temperatures of both and the tip position of a probe needle displaced by thermal deformation of both. The prober described in Patent Document 4 can bring the probe needle into efficient and stable contact with a semiconductor chip. [Prior art documents] [Patent documents]

[0010] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-173206 [Patent Document 2] Japanese Patent Application Laid-Open No. 2005-228788 [Patent Document 3] Japanese Patent Application Laid-Open No. 2007-311389 [Patent Document 4] Japanese Patent Application Laid-Open No. 2018-117095 Summary of the Invention [Problem to be solved by the invention]

[0011] In the probers described in Patent Documents 3 and 4, a prediction model is generated in advance. However, if the prediction model is generated when the temperature inside the prober is not stable, such as immediately after the temperature of the wafer chuck changes, a discrepancy occurs between the predicted value of the probe needle tip position by the prediction model and the actual measured value of the probe needle tip position. As a result, there is a risk of probing errors occurring. Therefore, in this case, it is necessary to continue generating (learning) the prediction model for a long time until the temperature inside the prober stabilizes.

[0012] Furthermore, even if a prediction model is used that was generated when the temperature inside the prober was stable, if drift in the temperature sensor occurs over a long period of time, or if the tip position of the probe needle shifts due to temperature fluctuations in areas of the prober where the temperature is not measured, there is a risk of a discrepancy occurring between the predicted value and the actual measured value of the tip position of the probe needle.

[0013] The present invention has been made in view of the above circumstances, and has as its object to provide a prober control device, a prober control method, and a prober that are capable of more accurately predicting the tip position of a probe needle. [Means for solving the problem]

[0014] A prober control device for achieving the object of the present invention is a prober control device that drives the relative movement unit of a prober that includes a wafer chuck that holds a wafer on which a plurality of semiconductor chips are formed, a probe card having probe needles, a card holder that holds the outer periphery of the probe card and faces the probe card to the wafer, and a relative movement unit that moves the wafer chuck relative to the probe needles, and brings the probe needles into contact with the semiconductor chips.The prober control device is also equipped with an input data acquisition unit that acquires input data including temperature data of at least one of the probe card and the card holder, a prediction unit that predicts the tip position of the probe needle based on the input data acquired by the input data acquisition unit using a prediction model that takes the input data as input and outputs the tip position of the probe needle, and a decision unit that decides whether or not to perform a prediction by the prediction unit based on the input data used as training data for machine learning of the prediction model before the prediction by the prediction unit and the input data acquired by the input data acquisition unit.

[0015] According to this prober control device, before the prediction by the prediction unit, it can be determined whether or not the prediction model used for this prediction can accurately predict the tip position of the probe needle based on the current input data.

[0016] In another aspect of the prober control device of the present invention, the determination unit performs the following processes: for each parameter of the input data, calculates the difference between the input data acquired by the input data acquisition unit and the input data used as training data; and calculates the square root of the sum of the squares of the differences for each parameter, and determines whether or not the prediction unit can perform a prediction based on whether or not there is at least one square root of the sum of the squares of the differences for each parameter that falls within a predetermined range.

[0017] In another aspect of the present invention, the prober control device includes a needle position acquisition unit that acquires the tip position of the probe needle when the decision unit determines that the result is negative, and a re-learning unit that re-learns the prediction model using teacher data that combines the input data acquired by the input data acquisition unit and the probe needle tip position acquired by the needle position acquisition unit, and the needle position acquisition unit, the re-learning unit, the input data acquisition unit, and the decision unit repeatedly operate until the decision unit determines that the result is positive. This allows the probe needle tip position to be accurately predicted.

[0018] In another aspect of the present invention, the prober control device includes a re-learning unit that removes the oldest input data and the probe needle tip position corresponding to the input data from the teacher data, and then re-learns the prediction model based on the teacher data, thereby reducing the influence of drift in the input data acquisition unit (temperature sensor).

[0019] In another aspect of the present invention, the prober control device includes a movement control unit that, when the determination unit determines that the probe needle is capable of being inserted, a prediction unit predicts the tip position of the probe needle, and controls the relative movement unit based on the tip position of the probe needle predicted by the prediction unit to bring the probe needle into contact with the semiconductor chip, thereby allowing the probe needle to be properly brought into contact with the semiconductor chip.

[0020] In another aspect of the prober control device of the present invention, the input data acquisition unit acquires, as input data, in addition to temperature data, alignment data including at least one of the chip size of the semiconductor chip, the position of the wafer, and the positional relationship between a first camera used to detect the semiconductor chip and a second camera used to detect the probe needle.

[0021] A prober for achieving the object of the present invention comprises a wafer chuck for holding a wafer on which a plurality of semiconductor chips are formed, a probe card having probe needles, a card holder for holding the outer periphery of the probe card and facing the probe card to the wafer, a relative movement unit for moving the wafer chuck relative to the probe needles, and the above-mentioned prober control device.

[0022] A prober control method for achieving the object of the present invention is a prober control method that drives the relative movement part of a prober that includes a wafer chuck that holds a wafer on which a plurality of semiconductor chips are formed, a probe card having probe needles, a card holder that holds the outer periphery of the probe card and faces the probe card to the wafer, and a relative movement part that moves the wafer chuck relative to the probe needles, and brings the probe needles into contact with the semiconductor chips. The prober control method includes an input data acquisition step that acquires input data including temperature data of at least one of the probe card and the card holder, a prediction step that predicts the tip position of the probe needle based on the input data acquired in the input data acquisition step using a prediction model that takes the input data as input and outputs the tip position of the probe needle, and a decision step that decides whether to execute the prediction step based on the input data used as training data for machine learning of the prediction model and the input data acquired in the input data acquisition step before the prediction step. [Effects of the Invention]

[0023] The present invention can more accurately predict the position of the probe needle tip. [Brief explanation of the drawings]

[0024] [Figure 1] 1 is a schematic diagram of a prober used in a wafer test system that inspects electrical characteristics of a plurality of semiconductor chips formed on a wafer. [Figure 2] FIG. 2 is a perspective view of the appearance of the prober; [Figure 3] FIG. 2 is a top view of a wafer held by a wafer chuck. [Figure 4] 10 is an explanatory diagram showing an example of temperature measurement points of a card holder and a probe card by a temperature sensor. FIG. [Figure 5] FIG. 2 is a functional block diagram showing functions of a control unit of the prober. [Figure 6] FIG. 10 is an explanatory diagram showing an example of training data used in machine learning of a prediction model by a prediction model generation unit. [Figure 7] 10 is a flowchart showing the flow of a method for contacting a probe needle with a semiconductor chip by a prober. [Figure 8] In a comparative example in which the determination unit does not perform determination and re-learning of the prediction model, a graph (see symbol VIIIA) showing the predicted value and actual measurement value of the tip position of the probe needle and a graph (see symbol VIIIB) showing the difference value between the predicted value and actual measurement value of the tip position of the probe needle are shown. [Figure 9] In this embodiment, there is a graph (see symbol IXA) showing the predicted and actual measured values ​​of the tip position of the probe needle, and a graph (see symbol IXB) showing the difference between the predicted and actual measured values ​​of the tip position of the probe needle. DETAILED DESCRIPTION OF THE INVENTION

[0025] [Prober configuration] Fig. 1 is a schematic diagram of a prober 10 used in a wafer test system that inspects the electrical characteristics of a plurality of semiconductor chips 9 (see Fig. 3) formed on a wafer W. Fig. 2 is a perspective view of the exterior of the prober 10.

[0026] 1 and 2, the prober 10 includes a base 12, a Y stage 13, a Y moving section 14, an X stage 15, an X moving section 16, a Zθ stage 17, a Zθ moving section 18, a wafer chuck 20, a support 23 (see FIG. 2), a head stage 24 (see FIG. 2), a card holder 25, a probe card 26, a wafer alignment camera 29, an upper and lower stage 30, a needle alignment camera 31, a cleaning plate 32, and a temperature sensor 34. The external configuration of the prober 10 is not limited to the example shown in FIGS. 1 and 2, and can be modified as appropriate.

[0027] A Y stage 13 is supported on the upper surface of the base 12 via a Y moving part 14 so as to be movable in the Y-axis direction.

[0028] Y moving unit 14 includes, for example, a guide rail that is provided on the upper surface of base 12 and is parallel to the Y axis, a slider that is provided on the lower surface of Y stage 13 and engages with the guide rail, and an actuator such as a motor that moves Y stage 13 in the Y axis direction. This Y moving unit 14 moves Y stage 13 on base 12 in the Y axis direction.

[0029] X stage 15 is supported on the upper surface of Y stage 13 via X moving unit 16 so as to be movable in the X-axis direction. X moving unit 16 includes, for example, a guide rail provided on the upper surface of Y stage 13 and parallel to the X-axis, a slider provided on the lower surface of X stage 15 and engaging with the guide rail, and an actuator such as a motor that moves X stage 15 in the X-axis direction. X moving unit 16 moves X stage 15 on Y stage 13 in the X-axis direction.

[0030] A Zθ stage 17 and a vertical stage 30 are provided on the upper surface of the X stage 15. A Zθ moving part 18 is provided on the Zθ stage 17. A wafer chuck 20 is held on the upper surface of the Zθ stage 17 via the Zθ moving part 18.

[0031] The Zθ moving unit 18 has, for example, an elevation mechanism that moves the Zθ stage 17 in the Z-axis direction (up and down direction), and a rotation mechanism that rotates the Zθ stage 17 around the Z-axis. Therefore, the Zθ moving unit 18 moves the wafer chuck 20 held on the upper surface of the Zθ stage 17 in the Z-axis direction and rotates it around the Z-axis.

[0032] The wafer W is held on the upper surface of the wafer chuck 20 by various holding methods such as vacuum suction. The wafer chuck 20 is also provided with a temperature adjustment unit 20a for adjusting the temperature of the wafer W. This temperature adjustment unit 20a may be a known mechanism such as a heater mechanism, a chiller mechanism, or a heat pump mechanism. The temperature adjustment unit 20a adjusts the temperature of the wafer W held on the wafer chuck 20.

[0033] The wafer chuck 20 is supported so as to be movable in the X, Y, and Z directions and rotatable about the Z axis via the already-described Y stage 13, Y moving unit 14, X stage 15, X moving unit 16, Zθ stage 17, and Zθ moving unit 18. This allows relative movement between the wafer W held on the wafer chuck 20 and probe needles 35, which will be described later. In other words, the Y stage 13 and Y moving unit 14, the X stage 15 and X moving unit 16, and the Zθ stage 17 and Zθ moving unit 18 function as relative movement units of the present invention.

[0034] Fig. 3 is a top view of the wafer W held by the wafer chuck 20. As shown in Fig. 3, a plurality of semiconductor chips 9 are formed on the wafer W. Each semiconductor chip 9 also has a plurality of electrode pads 9a formed thereon.

[0035] 1 and 2, the support pillars 23 are provided on the upper surface of the base 12 and support the head stage 24 above the Y stage 13, X stage 15, and Zθ stage 17 (hereinafter simply referred to as the stages 13, 15, and 17). This allows the head stage 24 to be fixed onto the base 12 via the support pillars 23.

[0036] A card holder 25 is held in the center of the head stage 24. A holding hole 25a that holds the outer periphery of a probe card 26 is formed in the card holder 25, and the probe card 26 is held in this holding hole 25a. As a result, the probe card 26 is held in a position facing the wafer W via the head stage 24 and the card holder 25.

[0037] The probe card 26 has probe needles 35 arranged according to the arrangement of the electrode pads 9a of the semiconductor chip 9 to be tested. The card holder 25 and the probe card 26 are replaced depending on the type of semiconductor chip 9.

[0038] The probe card 26 is provided with connection terminals (not shown) electrically connected to the probe needles 35, and a tester (not shown) is connected to these connection terminals. The tester supplies various test signals to the electrode pads 9a of the semiconductor chip 9 via the connection terminals of the probe card 26 and the probe needles 35, and receives and analyzes signals output from the electrode pads 9a to test whether the semiconductor chip 9 operates normally. Note that the configuration of the tester and the test method are well known technologies, so a detailed description thereof will be omitted.

[0039] The wafer alignment camera 29 corresponds to the first camera of the present invention, and photographs the semiconductor chip 9 of the wafer W held by the wafer chuck 20. Based on the photographed image taken by this wafer alignment camera 29, the positions of the electrode pads 9a of the semiconductor chip 9 to be inspected can be detected. Note that the installation position and structure of the wafer alignment camera 29 are not particularly limited, but in this embodiment, as disclosed in Japanese Patent Laid-Open Publication No. 2003-303865, an installation position and structure (spot light irradiation optical system) that can measure the relative distance to a needle alignment camera 31 (described later) are adopted.

[0040] A needle positioning camera 31 and a cleaning plate 32 are provided on the vertical stage 30 at positions substantially opposite the head stage 24, etc. The vertical stage 30 also has an elevation mechanism (not shown) that is movable in the Z-axis direction, and the Z-axis positions of the needle positioning camera 31 and the cleaning plate 32 can be adjusted. The needle positioning camera 31 and the cleaning plate 32 are supported movably in the X, Y, and Z axes via the Y stage 13 and Y moving unit 14, the X stage 15 and X moving unit 16, and the vertical stage 30. This allows the needle positioning camera 31, the cleaning plate 32, and the probe needle 35 to move relatively to each other.

[0041] The needle positioning camera 31 corresponds to the second camera of the present invention, and photographs the probe needle 35 of the probe card 26. The position of the probe needle 35 can be detected based on the photographed image of the probe needle 35 photographed by the needle positioning camera 31. Specifically, the XY coordinates of the tip position of the probe needle 35 are detected from the position coordinates of the needle positioning camera 31, and the Z coordinate of the tip position of the probe needle 35 is detected from the focal position of the needle positioning camera 31.

[0042] When inspecting semiconductor chips 9 on wafer W using prober 10 configured as described above, each time probe card 26 is replaced or a predetermined number of semiconductor chips 9 are inspected, stages 13, 15, 17 are driven to relatively move needle positioning camera 31 to a photographing position of probe needle 35, and then needle positioning camera 31 photographs probe needle 35. Based on the photographed image of needle positioning camera 31, the tip positions of probe needles 35 are detected as described above.

[0043] Furthermore, with the wafer W to be inspected held on the wafer chuck 20, the stages 13, 15, and 17 are driven to move the wafer alignment camera 29 relatively to the photographing position of the wafer W, and then the wafer alignment camera 29 photographs the semiconductor chip 9 on the wafer W. Based on the photographed image by the wafer alignment camera 29, the positions of the electrode pads 9a of the semiconductor chip 9 to be inspected are detected.

[0044] Then, the stages 13, 15, and 17 are driven to electrically contact the probe needles 35 with the electrode pads 9a of the semiconductor chip 9 to be inspected first. In this state, the semiconductor chip 9 to be inspected first is inspected by a tester (not shown). The remaining semiconductor chips 9 to be inspected are inspected in the same manner. Note that the specific method for inspecting the semiconductor chip 9 is well known, and therefore a detailed description thereof will be omitted here (see, for example, Patent Document 4).

[0045] The temperature sensors 34 are provided at positions facing the lower surfaces of the card holder 25 and the probe card 26, for example, on the side surfaces of the Zθ stage 17 and the upper and lower stages 30. Therefore, each temperature sensor 34 is held by each stage 13, 15, 17, 30 so as to be movable relative to the card holder 25 and the probe card 26.

[0046] Temperature sensor 34 is, for example, a non-contact temperature sensor that uses a radiation energy detection method, and measures the temperatures of card holder 25 and probe card 26 without contact. Card holder 25 and probe card 26 are thermally deformed by the temperature of wafer chuck 20, and this thermal deformation displaces the tip positions of probe needles 35. Therefore, by measuring the temperatures of card holder 25 and probe card 26 with temperature sensor 34, it is possible to predict the tip positions [displacement (displacement direction, displacement amount)] of probe needles 35 (see Patent Document 4 above).

[0047] Fig. 4 is an explanatory diagram showing an example of temperature measurement points of the card holder 25 and the probe card 26 by the temperature sensor 34. Note that the probe needles 35 are not shown in Fig. 4. As shown in Fig. 4, the temperature sensor 34 measures the temperatures at multiple locations on both the card holder 25 and the probe card 26, including multiple temperature measurement points P1 to P5 in the probe card 26 and multiple temperature measurement points P6 to P13 in the card holder 25, in order to detect the temperature distribution of both. Note that the temperature measurement points P1 to P13 in Fig. 4 are merely examples, and their positions and number may be changed as appropriate.

[0048] The temperature sensor 34 measures the temperature of each of the temperature measurement points P1 to P13 under the control of a control unit 40 (see FIG. 5), which will be described later, and outputs the temperature measurement results, that is, temperature data, to the control unit 40. When measuring the temperature of each of the temperature measurement points P1 to P13, the stages 13, 15, 17, and 30 are driven under the control of the control unit 40, which will be described later, so that the temperature sensor 34 is positioned so that it can measure the temperature of each of the temperature measurement points P1 to P13; that is, the temperature sensor 34 is moved relative to the card holder 25 and the probe card 26. This enables fixed-point measurement of the temperature of each of the temperature measurement points P1 to P13.

[0049] <Controller Functions> Fig. 5 is a functional block diagram showing the functions of the control unit 40 of the prober 10. Note that Fig. 5 shows only the functions of the control unit 40 relating to contact control between the probe needles 35 and the wafer W (electrode pads 9a of the semiconductor chip 9), and other functions are not shown because they are publicly known techniques.

[0050] 5, the control unit 40 corresponds to the prober control device of the present invention and controls all parts of the prober 10. The control unit 40 may be built into the main body of the prober 10, or may be provided separately from the main body.

[0051] The control unit 40 is configured by a computing device such as a personal computer, and includes a computing circuit configured with various processors, memories, etc. The various processors include a central processing unit (CPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), and a programmable logic device (e.g., simple programmable logic devices (SPLD), complex programmable logic devices (CPLD), and field programmable gate arrays (FPGA)). The various functions of the control unit 40 may be realized by a single processor, or by multiple processors of the same or different types.

[0052] In addition, the control unit 40 is connected to the aforementioned wafer alignment camera 29, needle alignment camera 31, temperature sensor 34, etc. via various communication interfaces (not shown), as well as to an alignment data measurement unit 38 and a memory unit 39.

[0053] The alignment data measurement unit 38 controls the wafer alignment camera 29, the needle alignment camera 31, etc. to measure alignment data. The alignment data is data used to predict the tip position (displacement) of the probe needle 35, along with the temperature data described above. This alignment data includes, for example, the three-dimensional chip size of the semiconductor chip 9 to be inspected, the three-dimensional position of the wafer W, and the three-dimensional relative distance between the wafer alignment camera 29 and the needle alignment camera 31 (hereinafter abbreviated as the camera relative distance). The camera relative distance indicates the positional relationship between the wafer alignment camera 29 and the needle alignment camera 31.

[0054] Specifically, the alignment data measurement unit 38 measures the chip size (amount of expansion) of the semiconductor chip 9 based on an image of the wafer W (semiconductor chip 9) captured by the wafer alignment camera 29. The alignment data measurement unit 38 also measures the position of the wafer W based on an image of a specific pattern (not shown) of the semiconductor chip 9 captured by the wafer alignment camera 29. Furthermore, as disclosed in Japanese Patent Application Laid-Open No. 2003-303865, the alignment data measurement unit 38 measures the camera relative distance using the wafer alignment camera 29, the needle alignment camera 31, and an optical system (not shown) that irradiates spot light. The alignment data measurement unit 38 then outputs alignment data including the chip size of the semiconductor chip 9, the position of the wafer W, and the camera relative distance to the control unit 40.

[0055] The memory unit 39 stores a control program (not shown) that operates the control unit 40, the inspection results of the semiconductor chip 9 by the prober 10, as well as teacher data 56 (also called training data) used in the machine learning of the prediction model 47 described below.

[0056] When inspecting the semiconductor chip 9 to be inspected within the wafer W, the control unit 40 executes a control program (not shown) read from the memory unit 39, thereby functioning as an input data acquisition unit 42, a needle position acquisition unit 44, a prediction unit 46, a prediction model generation unit 48, a determination unit 50, and a movement control unit 52.

[0057] Before contact control (hereinafter simply referred to as before contact control) in which probe needles 35 are brought into contact with semiconductor chip 9 to be inspected, and before generating and relearning a prediction model 47 described below, input data acquisition unit 42 executes temperature measurements at each of temperature measurement points P1 to P13 by temperature sensor 34 and alignment data measurements by alignment data measurement unit 38. As a result, input data acquisition unit 42 acquires input data including the temperature data at each of temperature measurement points P1 to P13 from temperature sensor 34 and alignment data from alignment data measurement unit 38 at each of the above-mentioned timings.

[0058] In addition, the input data acquisition unit 42 outputs the input data acquired before contact control to the prediction unit 46 and the determination unit 50 described below, and outputs the input data acquired before generating the prediction model 47 and before relearning to the prediction model generation unit 48 described below.

[0059] The needle position acquisition unit 44 causes the needle positioning camera 31 to photograph the probe needle 35 after replacing the probe card 26, after inspecting a predetermined number of semiconductor chips 9, and before generating and relearning the predictive model 47 described below, and acquires a photographed image of the probe needle 35 from the needle positioning camera 31, and acquires the tip position of the probe needle 35 based on this photographed image.

[0060] In addition, the needle position acquisition unit 44 outputs the tip position of the probe needle 35 acquired before generating the prediction model 47 and before relearning to the prediction model generation unit 48 described below, and outputs the tip position of the probe needle 35 acquired after replacing the probe card 26, etc. to the movement control unit 52 described below.

[0061] Prediction unit 46 predicts the tip position of probe needle 35 before contact control and when decision unit 50 (described later) has determined that execution of prediction by prediction unit 46 is "allowed." Specifically, prediction unit 46 predicts the tip position of probe needle 35 based on input data (temperature data and alignment data at temperature measurement points P1 to P13) acquired by input data acquisition unit 42, with reference to a prediction model 47 (described later) generated in advance, and outputs the tip position prediction result to movement control unit 52. Note that the tip position of probe needle 35 predicted by prediction model 47 also includes the amount of deviation (correction amount) from the tip position of probe needle 35 acquired by needle position acquisition unit 44.

[0062] Prediction model 47 is a trained model generated by machine learning (supervised learning) using a multiple regression model (multiple regression equation, also called multiple regression analysis) by prediction model generation unit 48 (described later). Prediction model 47 receives input of a plurality of input data (temperature data and alignment data for each of temperature measurement points P1 to P13) that are explanatory variables, and outputs a predicted value of the tip position of probe needle 35, which is a response variable.

[0063] The prediction model generation unit 48 generates a prediction model 47 before inspecting the semiconductor chips 9 on a product wafer W. First, the prediction model generation unit 48 measures teacher data 56 (input data, alignment data, and tip positions of the probe needles 35) using the product wafer W or an identical test wafer W (for creating a prediction model).

[0064] 6 is an explanatory diagram showing an example of training data 56 used in machine learning of prediction model 47 by prediction model generation unit 48. In order to prevent the drawing from becoming too complicated, Fig. 6 illustrates only alignment data in one direction (here, the Y direction) among the X, Y, and Z directions, and the tip position of probe needle 35 in one direction.

[0065] 6 and the previously described FIG. 5, the prediction model generation unit 48 controls the temperature sensor 34, the alignment data measurement unit 38, and the movement control unit 52 (described later) to measure input data (temperature measurement at each temperature measurement point P1-P13, alignment data measurement) for a predetermined time. This provides temperature data T1-T13 for each temperature measurement point P1-P13. Furthermore, alignment data includes the chip size D1 of the semiconductor chip 9 (amount of change from the start of the lot), the wafer position D2 of the wafer W (amount of change from the start of the lot), and the camera relative distance D3 (amount of change from the start of the lot).

[0066] The prediction model generation unit 48 controls the needle positioning camera 31, the needle position acquisition unit 44, the movement control unit 52 (described later), etc. in accordance with the measurement timing of the input data described above to measure the tip position of the probe needle 35. This allows the tip position Y[α] (α is an arbitrary natural number) of the probe needle 35 to be obtained for each measurement timing of the input data.

[0067] In this way, the prediction model generation unit 48 acquires a plurality of pieces of teacher data 56 including the input data and the tip positions Y[α] of the probe needles 35 corresponding to the input data. Note that, as will be described in detail later, in this embodiment, since re-learning of the prediction model 47 is possible, it is sufficient to acquire the minimum number of pieces of teacher data 56 necessary for machine learning of the prediction model 47; for example, in this embodiment, the teacher data 56 is acquired using one wafer W.

[0068] Next, the prediction model generation unit 48 generates a prediction model 47 for predicting the tip position of the probe needle 35 from the input data by machine learning using a multiple regression model, based on a plurality of teacher data 56, i.e., the input data (T1 to T13, D1 to D3) which are explanatory variables and the tip position Y[α] of the probe needle 35 which is a response variable. The specific method for generating this prediction model 47, i.e., the machine learning algorithm using the multiple regression model, is a well-known technique, so a detailed description will be omitted here. This enables the prediction unit 46 to predict the current tip position of the probe needle 35 from the current input data.

[0069] The machine learning algorithm for generating the prediction model 47 is not limited to a multiple regression model, and a known machine learning algorithm such as a convolutional neural network (CNN) may be used.

[0070] Furthermore, the prediction model generation unit 48 stores the training data 56 (or just the input data) used in the machine learning of the prediction model 47 in the storage unit 39. The training data 56 stored in the storage unit 39 is used by the determination unit 50, which will be described later, to determine whether or not to execute prediction by the prediction unit 46.

[0071] Furthermore, the prediction model generation unit 48, which will be described in detail later, operates after the prediction model 47 has been generated and when the decision unit 50, described later, decides that the prediction by the prediction unit 46 should not be performed, and re-learns the prediction model 47.

[0072] Returning to Figure 5, the decision unit 50 operates before contact control (before prediction by the prediction unit 46) and determines whether or not to execute prediction by the prediction unit 46 (simply referred to as determining whether or not to execute prediction) by comparing the current (latest) input data acquired by the input data acquisition unit 42 with the input data of the teacher data 56 in the memory unit 39.

[0073] If the current input data is not significantly different from the input data of teacher data 56 used in the machine learning of prediction model 47, prediction model 47 is in a trained state where machine learning has already been completed using teacher data 56 corresponding to the current input data. Therefore, when prediction unit 46 predicts the tip position of probe needle 35 using prediction model 47 based on the current input data, it can accurately predict the tip position of probe needle 35.

[0074] On the other hand, if the current input data deviates significantly from the input data of teacher data 56, prediction model 47 is in an unlearned state where machine learning has not been performed using teacher data 56 corresponding to the current input data. Therefore, even if prediction unit 46 predicts the tip position of probe needle 35 using prediction model 47 in an unlearned state, it will not be able to accurately predict the tip position of probe needle 35.

[0075] Therefore, the determination unit 50 compares the current input data with the input data of the teacher data 56 in the memory unit 39, and determines whether the prediction model 47 is in a learned state or an unlearned state for the current input data, thereby determining whether or not to make a prediction.

[0076] First, the determination unit 50 calculates the difference between the current input data and the input data of the teacher data 56 for each parameter of the input data (temperature data T1 to T13, chip size D1, wafer position D2, and camera relative distance D3). Next, the determination unit 50 calculates the square root of the sum of squares of the difference for each parameter, and determines whether the prediction model 47 is in a trained state or an untrained state based on whether there is at least one square root of the sum of squares for each parameter that falls within a certain range (below a threshold).

[0077] The following is a specific description of the determination method (judgment method) by the determination unit 50. To avoid complicating the description, the description will be given assuming that the input data consists of only temperature data T1 to T13.

[0078] When the number of parameters of the input data is m and the number of times learning has been completed is N, the learned input data (explanatory variables) is expressed by the following formula [1]. Also, the learned tip position Y[α] (objective variable) of probe needle 35 is expressed by the following formula [2]. And the function obtained by the machine learning algorithm using the multiple regression model, i.e., prediction model 47, is expressed by the following formula [3].

[0079]

number

[0080]

number

[0081]

number

[0082] If the "current input data" acquired at the stage of predicting the tip position of probe needle 35 is represented by X[T] shown in the following [Equation 4], the Euclidean distance D[s] between this current input data X[T] and X[s], which is the input data of the teacher data 56 at any sth time, is expressed by the following [Equation 5].

[0083]

number

[0084]

number

[0085] The determination unit 50 determines a predetermined threshold D for all Euclidean distances D[s] (s=1, 2, . . . , N). th and compare it with the threshold D th If there is at least one Euclidean distance D[s] that is less than , it is determined that the prediction model 47 (Equation 3) is in a learned state, and it is decided to perform prediction using this prediction model 47. In this case, the prediction unit 46 described above predicts the tip position Y[T] of the probe needle 35, which is the objective variable, using the prediction model 47 shown in Equation 3, based on the current input data X[T], which is the explanatory variable.

[0086] On the other hand, the determination unit 50 determines the threshold D th If there is no Euclidean distance D[s] that is less than , it is determined that the prediction model 47 is in an unlearned state, and it is decided not to perform prediction using this prediction model 47. In this case, the prediction model generation unit 48 receives the determination result from the determination unit 50 and functions as the re-learning unit of the present invention, thereby re-learning the prediction model 47.

[0087] When re-learning the prediction model 47, the prediction model generation unit 48 controls the needle position acquisition unit 44, the movement control unit 52 (described later), etc. to acquire the tip position Y[T] of the probe needle 35, which is the objective variable corresponding to the current input data X[T]. At this time, the prediction model generation unit 48 may control the temperature sensor 34, the alignment data measurement unit 38, the movement control unit 52, etc. to re-measure the input data.

[0088] Next, the prediction model generation unit 48 adds the current input data X[T] and the tip position Y[T] of the probe needle 35 to the teacher data 56 in the memory unit 39 to create new teacher data 56. At this time, in order to reduce the influence of drift of the temperature sensor 34, the prediction model generation unit 48 preferably excludes the oldest data (X[1], Y[1]) from the teacher data 56 (explanatory variables, objective variables) in the memory unit 39, as shown in the following [Equation 6] and [Equation 7].

[0089]

number

[0090]

number

[0091] Then, the prediction model generation unit 48 performs machine learning using a multiple regression model based on the teacher data 56 stored in the storage unit 39, i.e., the input data (explanatory variables) shown in the above formula [6] and the tip position of the probe needle 35 (objective variable) shown in the above formula [7], to re-learn the prediction model 47. As a result, a new prediction model 47 (function) is obtained as shown in the following formula [8].

[0092]

number

[0093] When the re-learning of the prediction model 47 is completed, the input data acquisition unit 42 acquires input data, and then the determination unit 50 determines whether or not the prediction is possible.

[0094] Thereafter, until the determination unit 50 determines that the prediction is possible, the needle position acquisition unit 44, the prediction model generation unit 48, the input data acquisition unit 42, and the determination unit 50 repeatedly operate to repeatedly acquire the tip position of the probe needle 35, update the teacher data 56 in the memory unit 39, re-learn the prediction model 47, acquire the input data, and determine whether the prediction is possible. This allows the prediction unit 46 to always predict the tip position of the probe needle 35 using the prediction model 47 in a trained state.

[0095] The movement control unit 52 drives the stages 13, 15, and 17 via the Y movement unit 14, the X movement unit 16, and the Zθ movement unit 18. The movement control unit 52 acquires the position of the semiconductor chip 9 (electrode pad 9a) to be inspected on the wafer W held on the wafer chuck 20 based on the captured image input from the wafer alignment camera 29. The movement control unit 52 also acquires the tip positions of the probe needles 35 (values ​​measured when the probe card 26 is replaced, etc.) from the needle position acquisition unit 44.

[0096] During inspection of a product wafer W, the movement control unit 52 drives each stage 13, 15, and 17 to move the wafer W relative to the probe needles 35, thereby bringing the probe needles 35 into contact with the semiconductor chips 9 to be inspected on the wafer W in order. At this time, the movement control unit 52 drives each stage 13, 15, and 17 based on the prediction result of the tip positions of the probe needles 35 by the prediction unit 46, and corrects the contact positions of the probe needles 35 with respect to the semiconductor chip 9 to be inspected for each semiconductor chip 9 to be inspected. As a result, even if the tip positions of the probe needles 35 are displaced due to thermal deformation of the card holder 25 and the probe card 26, etc., the probe needles 35 can be brought into contact with the semiconductor chips 9 to be inspected at each corrected contact position corresponding to the tip position after this displacement.

[0097] [Operation of this embodiment] 7 is a flowchart showing the flow of a method for contacting probe needles 35 with semiconductor chips 9 by prober 10 configured as described above, which corresponds to the prober control method of the present invention. Note that the explanation will be given assuming that prediction model 47 has been generated in advance, that training data 56 used in the machine learning of the prediction model 47 has been stored in storage unit 39, and that the tip positions of probe needles 35 have also been acquired by needle position acquisition unit 44.

[0098] When the product wafer W is held by the wafer chuck 20, the wafer alignment camera 29 photographs the semiconductor chip 9 of the wafer W. Then, the movement control unit 52 determines the position of the semiconductor chip 9 (electrode pad 9a) to be inspected based on the photographed image taken by the wafer alignment camera 29.

[0099] Furthermore, the input data acquisition unit 42 executes temperature measurements at each of the temperature measurement points P1 to P13 using the temperature sensor 34 and alignment data measurements using the alignment data measurement unit 38. As a result, the input data acquisition unit 42 acquires current input data including the temperature data at each of the temperature measurement points P1 to P13 and the alignment data (step S1, which corresponds to the input data acquisition step of the present invention).

[0100] When the acquisition of the current input data is completed, the determination unit 50 is activated to compare the current input data acquired by the input data acquisition unit 42 with the input data of the teacher data 56 in the storage unit 39 to determine whether or not the prediction is possible (step S2, which corresponds to the determination step of the present invention). Specifically, the determination unit 50 compares each of the Euclidean distances D[s] (s=1, 2, ..., N) expressed by the above [Equation 5] with the threshold D th Based on the result of the comparison, the prediction model 47 determines whether the current input data is in a learned state or an unlearned state.

[0101] If the determination unit 50 determines that the prediction is not possible (NO in step S3), the prediction model generation unit 48 controls the needle position acquisition unit 44, the movement control unit 52, etc. to acquire the tip position of the probe needle 35 corresponding to the current input data (step S4). Note that the input data may be reacquired at this time.

[0102] Next, as shown in the above formulas [6] and [7], the prediction model generation unit 48 adds the current input data and the tip position Y of the probe needle 35 to the teacher data 56 in the storage unit 39 and excludes the oldest data, thereby updating the teacher data 56 (step S5).The prediction model generation unit 48 then re-learns the prediction model 47 based on the new teacher data 56 in the storage unit 39 to generate a new prediction model 47 (step S6).

[0103] When the re-learning of the prediction model 47 is completed, the input data acquisition unit 42 acquires the input data again (step S2), and the determination unit 50 determines whether the prediction is possible based on the input data (step S3). Thereafter, the processes of steps S4 to S6, step S1, and step S2 are repeated until the determination unit 50 determines that the prediction is possible.

[0104] If the determination unit 50 determines that the prediction is possible (YES in step S3), the prediction unit 46 predicts the tip position of the probe needle 35 by referring to the prediction model 47 based on the input data acquired in the most recent step S1 (step S7, which corresponds to the prediction step of the present invention). Then, the prediction unit 46 outputs the prediction result of the tip position of the probe needle 35 to the movement control unit 52.

[0105] Next, movement control unit 52 controls the movement of each stage 13, 15, 17 based on the prediction result of the tip position of probe needle 35 input from prediction unit 46 and the previously determined position of semiconductor chip 9 to be inspected, to bring probe needle 35 into contact with semiconductor chip 9 to be inspected (step S8). After this contact, inspection of semiconductor chip 9 is carried out by a tester (not shown) (step S9).

[0106] Subsequently, the remaining semiconductor chips 9 to be inspected are inspected in the same manner. At this time, the processes from step S1 to step S7 may be repeatedly executed every time a predetermined number of semiconductor chips 9 are inspected or every time a predetermined time has elapsed.

[0107] [Effects of this embodiment] As described above, in this embodiment, the determination unit 50 compares the current input data acquired before contact control with the input data of the teacher data 56 to determine whether or not a prediction is possible, and if the determination unit 50 determines that a prediction is not possible, the prediction model 47 is re-learned, thereby making it possible to predict the tip position of the probe needle 35 more accurately than before.

[0108] Fig. 8 is a graph (see symbol VIIIA) showing the predicted value PV and actual measurement value MV of the tip position of the probe needle 35 in a comparative example in which determination by the determination unit 50 and re-learning of the prediction model 47 are not performed, and a graph (see symbol VIIIB) showing the difference between the predicted value PV and actual measurement value MV of the tip position of the probe needle 35. Fig. 9 is a graph (see symbol IXA) showing the predicted value PV and actual measurement value MV of the tip position of the probe needle 35 in this embodiment, and a graph (see symbol IXB) showing the difference between the predicted value PV and actual measurement value MV of the tip position of the probe needle 35.

[0109] 8 and 9 show the time changes in the predicted value PV and the measured value MV of the tip position of the probe needle 35 in any one direction (here, the Y direction) among the X, Y, and Z directions, and the time changes in the difference between them, when the temperature of the wafer chuck 20 is set to 200° C. Furthermore, the symbol WA in FIGS. 8 and 9 indicates the machine learning range in which machine learning was performed.

[0110] As shown in Figure 8, in the comparative example, drift of the temperature sensor 34 occurs, or the tip position of the probe needle 35 shifts due to temperature fluctuations in areas within the prober 10 where the temperature is not measured, and it was confirmed that a discrepancy occurs between the predicted value PV and the actual measured value MV of the tip position of the probe needle 35 after the machine learning range WA, and the difference value gradually increases.

[0111] 9, in this embodiment, even after the machine learning range WA, by determining whether or not a prediction is possible by the determining unit 50 and re-learning the prediction model 47, it was confirmed that the predicted value PV and the actual measured value MV of the tip position of the probe needle 35 almost agree with each other and the difference value is reduced. As a result, in this embodiment, even if drift occurs in the temperature sensor 34 or the tip position of the probe needle 35 is displaced due to temperature fluctuations at points in the prober 10 where the temperature is not measured, the re-learning of the prediction model 47 makes it possible to more accurately predict the tip position of the probe needle.

[0112] In addition, in this embodiment, by enabling the re-learning of the prediction model 47, it is no longer necessary to continue generating the initial prediction model 47 (machine learning) for a long period of time, thereby reducing the work required to generate the prediction model 47.

[0113] [others] In the above embodiment, the prediction model generation unit 48 performs both the generation and re-learning of the prediction model 47, but the generation of the prediction model 47 may be performed by the manufacturer of the prober 10 or another prober 10. In this case, a re-learning unit that only re-learns the prediction model 47 may be provided in the control unit 40 instead of the prediction model generation unit 48.

[0114] In the above embodiment, the determination unit 50 determines whether or not the prediction is possible using the above [Equation 5], but this determination method is not particularly limited. For example, for each parameter of the input data, it may be determined whether or not the current input data is included between the maximum and minimum values ​​of the input data of the teacher data 56 (hereinafter referred to as the maximum-minimum range), and the prediction may be determined based on whether or not the current input data is included in the maximum-minimum range for all parameters.

[0115] In the above embodiment, the temperature data of the card holder 25 and the probe card 26 and the alignment data are measured as input data, but only the temperature data may be measured. Also, in the above embodiment, the temperature data of both the card holder 25 and the probe card 26 are measured as input data, but the temperature data of at least one of the card holder 25 and the probe card 26 may be measured. Furthermore, in the above embodiment, the chip size of the semiconductor chip 9, the position of the wafer W, and the camera relative distance are measured as alignment data, but at least one of these may be measured.

[0116] In the above embodiment, a non-contact type temperature sensor 34 is used, but a contact type temperature sensor 34 may also be used. [Explanation of symbols]

[0117] 9...semiconductor chip, 9a...electrode pad, 10...prober, 12...base, 13...Y stage, 14...Y moving part, 15...X stage, 16...X moving part, 17...Zθ stage, 18...Zθ moving part, 20...wafer chuck, 20a...temperature adjusting part, 23...support, 24...head stage, 25...card holder, 25a...holding hole, 26...probe card, 29...wafer alignment camera, 30...upper and lower stages, 31...needle alignment camera, 32...cleaning plate, 34...temperature sensor, 35...probe Lobe needle, 38... alignment data measurement unit, 39... memory unit, 40... control unit, 42... input data acquisition unit, 44... needle position acquisition unit, 46... prediction unit, 47... prediction model, 48... prediction model generation unit, 50... determination unit, 52... movement control unit, 56... training data, D... Euclidean distance, D1... chip size, D2... wafer position, D3... camera relative distance, Dth... threshold, MV... actual measurement value, P1 to P13... temperature measurement points, PV... predicted value, T1 to T13... temperature data, W... wafer, WA... machine learning range

Claims

1. A prober control device that drives a relative movement unit that relatively moves a probe card having probe needles and a card holder that faces the probe card to an object to be tested for electrical characteristics, thereby bringing the probe needles into contact with the object to be tested, an input data acquisition unit that acquires input data including temperature data of at least one of the probe card and the card holder; a prediction unit that predicts the tip position of the probe needle based on the input data acquired by the input data acquisition unit, using a prediction model that uses the input data as an input and the tip position of the probe needle as an output; a decision unit that decides whether or not to execute a prediction by the prediction unit based on the input data used as training data for machine learning of the prediction model before the prediction by the prediction unit and the input data acquired by the input data acquisition unit; A prober control device comprising:

2. A prober control method for driving a relative movement unit that relatively moves a probe card having probe needles and a card holder that positions the probe card facing an object of electrical characteristic testing, thereby bringing the probe needles into contact with the object of testing, comprising: an input data acquisition step of acquiring input data including temperature data of at least one of the probe card and the card holder; a prediction step of predicting the tip position of the probe needle based on the input data acquired in the input data acquisition step, using a prediction model that uses the input data as an input and the tip position of the probe needle as an output; a determination step of determining whether or not to execute the prediction step based on the input data used as training data for machine learning of the prediction model and the input data acquired in the input data acquisition step, prior to the prediction step; A prober control method comprising:

Images